innovation and technical change; (c) institutional changes leadin effective organization and management at both governmental a levels; and (d) a better environment for entrepreneurs. These j interrelated, and all depend to some extent on improvements in edi Recent Educational Expansion The number of children enrolled in the primary schools of the oped countries rose 150 percent during the 15 years from 1950 to the percentage of all children 6 to 12 years old who were in school less than 40 percent to more than 60 percent. This marked increas< ment ratios (the fraction of the total age group who are in school) r large measure the value placed on education by people of all c income groups in the developing countries. Public pressure for more education probably came in part creasing economic returns to skill and education as industrial ceeded and in part from the widening disparity between the i people who had some formal education and those who were illit disparity in turn came from the growing demand for skilled lab stagnation in demand for uneducated and unskilled workers, whos were rapidly increasing because of high rates of population growth, four Latin American countries and in India show that the earning: with 5 to 6 years of schooling are double or triple those of person; spent less than 2 years in school. Persons with 11 years of educ three to six times as much as functional illiterates. Education of Children as a Form of Saving and Investment Educational expansion means that many parents have been sper to improve the education and skills of children even though this h more difficult as the number of children in each family increa: investments in the "quality" of children may be taking place at tl of savings by households and corresponding capital investment in tl means of production. Statistical analysis of a large number of less countries shows that the level of savings measured in terms of n come remains, over time, a relatively constant fraction of per capit This fraction does not increase as per capita incomes rise, but froi to country it shows a strong inverse correlation with child depende that is, the proportion of children less than 15 years old to adult, years old.* (As we have seen, these dependency ratios depend c *Nathaniel H. Leff, "Dependency Rates and Savings Rates," Amer Econrated and learned in hew contexts and locations, new types of risks have to assumed, and new social and economic relationships have to be forged.